Preparation of dataset:

library(ggplot2)
library(plotly)

# Sets directory with .txt files for each subjects
mypath = "/home/finc/Dropbox/Copy/001_RESEARCH/001_PROJECTS/001_NOW/000_LearningBrain/2018/training_progress/raw"
setwd(mypath)

# Creates list of text files
txt_files_ls = list.files(path = mypath, pattern = "*.txt") 

# Reads the files (columns separated with tab)
txt_files_df <- lapply(txt_files_ls, function(x) {read.table(file = x, header = F, sep ="\t")})

# Combines all text file into one data frame
train_progress <- do.call("rbind", lapply(txt_files_df, as.data.frame)) 

# Adds column with subjects IDs and sesion number
subjects <- substr(txt_files_ls, 1, 5) # cuts subjects' ids 
id <- cbind(rep(subjects, each = 18)) # multiplies them by number of session (same for each subject) and changes row vector to a column vector
session <- cbind(rep(c(1:18), times = 19)) # creates column vector with numbers of sessions
training_progress <- cbind(id, session, train_progress) # adds IDs as a new colum

# Changes n-back level variables to numeric
for (i in 3:22){
 training_progress[, i] <- as.numeric(training_progress[, i])  
}

# Calculates mean of obtained n-back level during one training session session
training_progress$mean_level <- rowMeans(training_progress[, 3:22], na.rm = T)
training_progress$max_level <- do.call(pmax, training_progress[3:22])

  
# Summary of the data
# str(training_progress)
# summary(training_progress)

Mean n-back level for all subjects:

# Mean n-back level for all subjects, all sessions
lm(mean_level ~ session, data = training_progress)
## 
## Call:
## lm(formula = mean_level ~ session, data = training_progress)
## 
## Coefficients:
## (Intercept)      session  
##      2.2216       0.0727
# Scatter plot
g <- ggplot(training_progress,
              aes(x = session, y = mean_level))
g <- g + geom_point()
g <- g + geom_smooth(method = "lm")
g <- g + theme_bw()
g <- g + ggtitle("Mean n-back level for all subjects") 
g <- g + xlab("Session") 
g <- g + ylab("Mean n-back level") 
g <- g + theme_bw()

ggplotly()

Mean n-back level for each individual:

# Fits regression line to each individual's progress
slope_mean <- c()
intercept_mean <- c()

for (i in subjects){
  sub_data = training_progress[training_progress$id == i, ]
  line = lm(mean_level ~ session, data = sub_data)
  intercept_mean[i] <- line$coefficients[1]
  slope_mean[i] <- line$coefficients[2]
}

# Makes interactive plot with mean n-back level for each subject
g <- ggplot(training_progress,
              aes(x = session, y = mean_level))

g <- g + geom_line(data = training_progress,  
                       aes(x = session, y = mean_level, group = id),
                       colour = "gray") #, alpha = 1/2, size = 1/2)
g <- g + geom_smooth(method = "lm", size = 0.5)
g <- g + geom_line(size = 0.5)  
g <- g + facet_wrap(~ id)
g <- g + theme_bw()
# g <- g + ggtitle("Mean n-back level for each subject") 
g <- g + xlab("Session") 
g <- g + ylab("Mean n-back level") 

ggplotly()
# Histograms for intercept and slope
hist(intercept_mean, xlab = "Intercept")

hist(slope_mean, xlab = "Slope")

Maximum n-back level for all subjects:

# Maximum n-back level for all subjects, all sessions
lm(mean_level ~ session, data = training_progress)
## 
## Call:
## lm(formula = mean_level ~ session, data = training_progress)
## 
## Coefficients:
## (Intercept)      session  
##      2.2216       0.0727
# Scatter plot
g <- ggplot(training_progress,
              aes(x = session, y = max_level))
g <- g + geom_point()
g <- g + geom_smooth(method = "lm")
g <- g + theme_bw()
g <- g + ggtitle("Maximum n-back level for all subjects") 
g <- g + xlab("Session") 
g <- g + ylab("Maximum n-back level") 
g <- g + theme_bw()

ggplotly()

Maximum n-back level for each individual:

# Fits regression line to each individual's progress
slope_max <- c()
intercept_max <- c()

for (i in subjects){
  sub_data = training_progress[training_progress$id == i, ]
  line = lm(max_level ~ session, data = sub_data)
  intercept_max[i] <- line$coefficients[1]
  slope_max[i] <- line$coefficients[2]
}

# Makes interactive plot with mean n-back level for each subject
g <- ggplot(training_progress,
              aes(x = session, y = max_level))

g <- g + geom_line(data = training_progress,  
                       aes(x = session, y = max_level, group = id),
                       colour = "gray") #, alpha = 1/2, size = 1/2)
g <- g + geom_smooth(method = "lm", size = 0.5)
g <- g + geom_line(size = 0.5)  
g <- g + facet_wrap(~ id)
g <- g + theme_bw()
# g <- g + ggtitle("Maximum n-back level for each subject") 
g <- g + xlab("Session") 
g <- g + ylab("Maximum n-back level") 

ggplotly()
# Histograms for intercept and slope
hist(intercept_max, xlab = "Intercept")

hist(slope_max, xlab = "Slope")